1. Relative Cyclone Risk Was Assessed at Two Spatial Scales in Southeastern Bangladesh. 2. Conceptual Structure of G

Highlights: 1. Relative cyclone risk was assessed at two spatial scales in southeastern Bangladesh. 2. Conceptual structure of general risk model was brought to practice for the assessment. 3. Diverse data representing the cyclone hazard, exposure, and vulnerability was analyzed and integrated. 4. Complementary use of AHP and GIS has been valuable for projecting the cyclone risk. 5. A reasonable consistency was noticed between the simulated risk and experiential impacts. Cyclone risk assessment of the Cox’s Bazar district and Rohingya refugee camps in southeast Bangladesh Akhtar Alam1*, Peter Sammonds1, Bayes Ahmed1 1Institute for Risk and Disaster Reduction, University College London, Gower Street, London, WC1E 6BT, United Kingdom. *Corresponding author: AA ([email protected]) Abstract Bangladesh has a long history of devastating tropical cyclones. In view of the effects of the storms on the country, risk assessment is essential for devising the mitigation strategies at various levels. 1 By way of bringing the conceptual structure of general risk model in practice, this work aims to examine the spatial patterns of cyclone risk in the Cox’s Bazar district (I) and Rohingya refugee camps (II) located on the southeastern coast of Bangladesh. We use 14 parameters representing the hazard, exposure, and vulnerability as the components of risk. The selected parameters were analyzed and integrated though the complementary use of Analytic Hierarchy Process (AHP) and Geographic Information System (GIS) for depicting the cyclone risk situation comprehensively at both the spatial scales. The status of the cyclone risk was identified and quantified as very high (6.84%, 3.43%), high (45.78%, 27.82%), moderate (5.97%, 39.42%), low (40.62%, 28.70%), and very low (0.81%, 0.61%) for the spatial scale I and II respectively. In general, northwestern and southern peripheral areas exhibited higher risk than the central and northeastern parts of the Cox’s Bazar district; and in the refugee settlements, camp number 1E, 1W, 7, and 13 revealed relatively higher levels of the risk. The results of the assessment (I) were correlated with experiential damage from the 1991 cyclone; a reasonable consistency was noticed between the simulated scenario and the observed impacts. We assume that the deliverables of this spatial analysis could be useful to stakeholders while formulating the cyclone risk mitigation policies for the region. Furthermore, this work demonstrates that the applied method would deliver reliable results if tested in other coastal environments. Keywords: Bangladesh; Cyclone risk; General risk model; AHP and GIS; Cox’s Bazar; Rohingya refugees 1. Introduction 2 Tropical cyclones are characterized by high speed winds, extreme rainfall, and storm surge. These attributes often make them violent, resulting in colossal loss of life, widespread destruction of infrastructure, and emergence of diseases along the coastal areas of the world (Shultz et al., 2005; Hong and Möller, 2012; Krapivin et al., 2012; Mori and Takemi, 2016). Last 30–40 years have particularly seen an increase in strong cyclones around the world (Varotsos et al., 2015). The storms like Bhola (1970), Tracy (1974), Andrew (1992), BOB 06 (1999), Katrina (2005), Sidr (2007), Nargis 2008, Sandy (2012), Haiyan (2013), Hudhud (2014) Patricia (2015) and Idai (2019) are the recent examples of the tropical cyclones which caused enormous human causalities and economic loss (Willoughby and Black, 1996; Emanuel, 2005; Brunkard et al., 2008; Lin et al., 2009; Paul, 2009; Lagmay et al., 2015; Huang et al., 2017). Among the historical extreme storm events, the Bhola (1970) that killed 0.3-0.5 million people is considered as deadliest ever recorded cyclone and the hurricane Katrina that destroyed more than 200,000 homes and other infrastructure is the costliest with an economic loss of US$125 billion (Vigdor, 2008; Fritz et al., 2009; Peduzzi et al., 2012; Deryugina et al., 2014). However, the nature of risks posed and the magnitude of impacts varies considerably form one region to another (Resio and Irish, 2015). Between 1980 and 2009 cyclones affected 466 million people, resulting in death of 412,644 and injury to 290,654 with less developed nations in Asia experiencing the maximum mortality and injury (Doocy et al., 2013). On an average, a tropical storm landfall in the north Indian Ocean results in death of about 2000 persons which is much higher compared to average fatalities per landfall in any other ocean basin of the world (Seo and Bakkensen, 2016). Bangladesh is locus of hydrometeorological hazards. The country has been facing the brunt of the tropical cyclones mainly because of its location and lowland topography (Khalil, 1992; Ali, 1996; Alam et al., 2003; Shamsuddoha and Chowdhury, 2007; Alam and Collins, 2010; Haque 3 and Jahan, 2016). Bangladesh experiences cyclones almost each year during early summer and retreating rainy season; as a result, cyclone related deaths have been recorded as more than one million since 1877 (Paul and Dutt, 2010; Dasgupta et al., 2014). In fact, most of the world’s catastrophic cyclones have been those hitting Bangladesh e.g., the episodes of 1584, 1737, 1942, 1876, 1897 and 1970. The event of 29 April 1991 is one of the deadliest in the series; the storm struck the eastern coast of the country with wind speeds exceeding 240 km/h, generating storm surge of more than 9 meters above mean sea level, killing 138,000-145,000 people and resulting in economic loss of $2.07 billion (Ministry of Health and Family Welfare, 1992; Bern et al., 1993; Khalil, 1993; Ikeda, 1995). Another cyclone in 2007 (Sidr) caused death of 3,406 people (Paul, 2009) and economic loss of $1.67 billion (Dasgupta et al., 2010). The economically deprived and marginalized populations living along the densely populated coastal areas are the most affected and compelled to remain under the continuing threat of the storms because of landlessness (UNICEF, 1993). Not only is the large number of deaths a concern, the magnitude of the economic loss from the storms is also too high for an economy like Bangladesh. The economic burden is further aggravated by the rehabilitation costs, taking substantial share of the Gross Domestic Product (GDP) after every catastrophic cyclone event in the country. Moreover, succeeding the primary effects, the cyclones have also been causing sanitation issues and disease epidemics owing to scarcity of fresh water in the storm hit areas (Hoque et al., 1993). Although, Bangladesh is frequently effected by the tropical cyclones, the risk levels of different areas remain largely unknown. Periodic evaluation of the hazard severity, exposure and vulnerability conditions at varied spatial scales is imperative for recognizing the risk of coastal areas and alleviating the impact of future cyclones in the country. Recently, many Geographic Information System (GIS) based attempts have been made for assessing the cyclone hazard, 4 vulnerabilities, and mitigation capacities in Bangladesh (e.g., Rana et al., 2010; Hoque et al., 2016; Hoque et al., 2017; Hossain and Paul, 2017; Quader et al., 2017; Hoque et al., 2018; Hossain et al., 2019). The investigation by Hoque et al. (2019) that performs the cyclone risk assessment of eastern coast seems most relevant to the present work because of some spatial overlap in the area of interest (Cox’s Bazar district) and identical methods adopted (AHP and GIS). Even though there is variation among the choice of parameters and the interpretations thereof between the two studies, the present one provides an opportunity to make comparisons for understanding how the selection of parameters, weightage of the parameters and human bias can influence the results of multiple criteria decision making studies even with the similar objectives. Moreover, focus on the Rohingya refugees is also a uniqueness of the present analysis because the Rohingya humanitarian crisis gained global attention exclusively from the perspective of conflict and the risk posed by various natural hazards to about 1 million people as a result of 2017 exodus remains absolutely unspecified. Risk assessment has been recognized as a priority action for building the resilience of the communities and preventing the disasters (Sendai Framework, 2015). By combining the possible role of a hazard, exposure, and vulnerability, the risk assessment illustrates how a system is expected to be affected in future. Therefore, pre-event assessment is an opportunity to comprehend the status of risk in a particular area, initiate mitigation measures and reduce the anticipated losses. As a component of risk, hazard implies the nature, intensity, location, and frequency of a process (natural hazard); whereas exposure is spatial in context and describes the people and assets at a particular location with likelihood of being effected by the prevailing hazards (UNDP, 2010). Vulnerability on the other hand considers the physical, social, economic, political, and environmental characteristics of a community or a system as the fundamental features for the risk 5 assessment (Martine and Guzman, 2002; Turner et al., 2003; Adger 2006; Dwyer, 2004; Douglas, 2007). Vulnerability is vital for understanding the conditions that enables a hazard to become a disaster (Tapsell et al., 2010). Although vulnerability may seem to be an intuitively simple notion, it is complex to define and even more difficult to quantify and apply in practice (UNEP, 2002). Comprehensive vulnerability analysis considers the totality of the system; however, real world constrictions necessitate a ‘reduced’ vulnerability assessment (Turner et al., 2003). Geographic Information System (GIS) coupled with different statistical techniques provides an effective decision support environment for the multiple criteria based assessment of risks related to various natural hazards (e.g., Zerger and Smith, 2003; Gillespie et al., 2007; Rana et al., 2010; Alam et al., 2018; Bhat et al., 2018) including cyclones (Taramelli et al., 2008; Klemas, 2009; Ozcelik et al., 2012; Mahapatra et al., 2015, Hoque et al., 2018; Mansour, 2019; Nguyen et al., 2019).

Load more